TeloframeTeloframe
Product
The loopFor clubsPricingLibraryDrills
Sign inStart free
Feature · AI Coaching Assistant

Eight tools.
No chat box.

Teloframe has no assistant to talk to. It has eight buttons, each in the place you already do the work, each reading the record you already built. They write the paragraph you were going to write anyway: the progress report, the parent summary, the review prep, the opponent brief, the debrief. You keep every decision.

Start free →See all eight
teloframe.com / teams / u16 · AI match debrief
Match debriefOpponent prepProgress report
✓ SavedLast generated 2 hours ago · 1 credit
Debrief · from the logged record
OBSSecond-half drop in first contact after the double changeTIMELINE
PRITraining priority · restart speed from a dead ballHIGH
PRITraining priority · pressing trigger on the switchMEDIUM
PLRJordan M. · 62 mins · linked to goal: weak-foot releaseIDP
SETSet pieces · near-post delivery from the left, 3 of 5NOTED
Read from
◆ This match record only☷ Your last 4 session themes↑ Squad development goals✕ No external scouting data
Eight tools · each in its own screen
↑Progress report
☷Parent summary
◱Review prep brief
◉Opponent prep
◆Match debrief
▲Club standards check
◲Whiteboard photo import
☰Fixture text import
What went into this generation
MATCHvs Ashfield U16 · 2 to 2 · home
TIMELINE14 logged events · 18 player entries
SESSIONSLast 4 themes · pressing, restarts
GOALSActive development goals for the squad
MODELgpt-5.4 · fixed output shape
✦Eight tools · not one chat box◱Your record only · no external data✓Nothing auto-sends · every send is a click◎Priced in credits · a failed run is free
8tools
All live, all in the app today
1 to 4credits
Per generation, charged at cost
10free
Welcome credits, no card needed
Fixed shape
Schema-checked output, not loose prose
Zero
Credits for a repeat, or a run that failed
No chat
Nothing to prompt, nothing to jailbreak
01 · The eight tools

Every one of them writes something you were going to write anyway.

Eight buttons, not one assistant. Each has a fixed job, a fixed output shape and a fixed price. Six work on every plan, one needs Pro, one is for club library reviewers. None of them asks you what you want.

↑
Progress report
Player file · Overview
Reads the player's evaluations, active development goals, match appearances and season context and writes the progress account you would otherwise write from scratch before a handover.
WritesPeriod summary, dimension trends, goal progress, attendance, match performance, strengths, development areas, recommendations
LimitEach dimension average is overwritten with the figure the server computes from the record, so the report cannot invent a score.
2 credits held, 1 to 2 chargedEvery plan
☷
Parent summary
Player file · Plan tab
The same record rewritten for a parent, in plain language, with football terms explained rather than assumed. Strengths first, growth areas framed as work in progress.
WritesSix sections: greeting, strengths, areas for growth, recent highlights, next steps, closing
LimitSending is a separate click. It only reaches guardians already linked to the player portal, and the send is written to the workspace audit log.
1 creditEvery plan
◱
Review prep brief
Player file · Reviews tab
What changed since the last formal review, with counts and dates off the record, then one talking point and one question worth asking for each active goal.
WritesFive sections: since last review, evidence highlights, goal talking points, growth areas, suggested next actions
LimitRead only. It cannot fill in the review form, score a rubric or close a goal.
1 creditEvery plan
◉
Opponent prep
Team page · before a fixture
A pre-match brief built only from the fixtures you have already logged against this opponent: scores, both formations, your match notes and tactical notes, plus your recent session themes.
WritesHead to head line, key patterns, threats to watch, suggested focus areas, training recommendations, formation notes
LimitNo scouting feed, no external database. If you have played them once, it says what one match can support.
1 creditEvery plan
◆
Match debrief
Team page · after a fixture
Turns the logged result into next week's priorities: the timeline events, minutes and contributions you recorded, read against the active development goals of the players who were involved.
WritesSummary, key observations, training priorities, player highlights, set-piece analysis, formation notes
LimitSaved to the match record. Reopening loads the saved debrief with no second charge until you choose to regenerate.
1 creditEvery plan
▲
Club standards check
Club workspace · Standards tab
Scores a drill or session submitted to the club library against your own written methodology and age-band guidance, and names the standard each finding refers back to.
WritesScore out of 100, verdict, summary, up to 15 findings with severity and a suggested fix, strengths, age-band compliance
LimitThe score and verdict are computed from the findings by rule, not chosen by the model, so the same submission scores the same twice. Labelled AI estimate, advisory: a human reviewer still decides.
1 credit held, 1 to 2 chargedClub, library reviewers
◲
Whiteboard photo import
Tactics board · Import from photo
Photograph the whiteboard at the end of a session and get the same shape back as an editable board scene instead of retyping it.
WritesUp to 30 players with side, number and position, the ball, up to 20 arrows and circles, both team colours, a confidence score
LimitLow confidence refuses to apply and tells you what it could not read. Applying asks you to confirm before it replaces what is on the board.
3 credits held, 2 to 4 chargedEvery plan
☰
Fixture text import
Sessions and team pages · Import fixtures
Paste the league email, the message thread or the fixture page and it pulls out the fixtures. CSV and calendar files stay free on every plan, because fixture data should never be behind a paywall.
WritesFixture rows: date, kickoff, opponent, home or away, competition, venue, matchday, notes
LimitExtraction only. It is told never to invent a fixture or guess a kickoff time, and every row lands in a review table you edit before anything is saved.
2 credits held, 1 to 3 chargedPro and Club
02 · Where it shows up

Not an app you open. A button in the screen you were in.

There is no AI section in the navigation, because there is nothing to visit. Each tool sits in the screen that already holds its context, so the thing it needs to read is the thing you are looking at.

◆
Sunday night
Team page · the match you just logged
You have already entered the score, the timeline and the minutes. The debrief turns that into three or four training priorities for the week, each one pointed at a development goal a player in that squad is actually working on. It saves to the match, so Tuesday's you can read it again for nothing.
Match debrief
◲
Tuesday, 9pm
Tactics board · Import from photo
The shape you drew on the dressing-room whiteboard is on your phone. Photograph it, upload it, and the players, ball, arrows and circles come back as an editable board scene with both team colours picked out. If the photo is too dark to read, it says so instead of guessing.
Whiteboard photo import
◉
Thursday
Team page · the fixture ahead
A brief built from the times you have played them before: the scores, both formations, the notes you wrote at the time, set against the themes of your recent sessions. Nothing scraped, nothing bought in. If your history with them is thin, the brief is short and says why.
Opponent prep
↑
Review week
Player file · Reviews, Overview, Plan
Three separate buttons in three tabs, in the order you actually work: the prep brief before you sit down with the player, the progress report when you need the written account, the parent summary when the family meeting is next. Each is a draft you edit. None of them fills in the review form.
Review prep briefProgress reportParent summary
▲
When a coach submits
Club workspace · Standards tab
A drill lands in the club library queue. The check reads it against your written methodology and your age-band guidance and returns findings with severities and suggested fixes, so the reviewer starts from a list instead of a blank page. The reviewer still approves or rejects.
Club standards check
☰
Season start
Sessions and team pages · Import fixtures
The league sends a fixture list as an email, a message or a web page nobody can export. Paste it in and the dates, opponents, kickoffs and venues come out as rows you check before saving. A clean CSV or an .ics file needs no AI and no credits, on any plan.
Fixture text import
03 · What goes into the prompt

It reads your record. That is the whole context.

Each tool has its own loader that queries named tables in your workspace and builds the prompt from what comes back. There is no browsing step, no scouting feed and no shared corpus. If a fact is not in your data, it is not in the output.

The four things a loader can read
01
The player record
Evaluations and their dimension scores, the development plan and its active goals, reflections, appearances, minutes, goals and assists
Progress · parent · review prep
02
The match record
Result, formations, the event timeline you logged, per-player entries and notes, and the development goals of the players involved
Debrief · opponent prep
03
Your recent sessions
The themes of the sessions you have been running, so a brief points at work you are already doing
Debrief · opponent prep
04
Your club's written standards
The club profile, age-band guidance, standards templates and the drills and sessions your club has already approved
Standards check
Every one of those queries is filtered to the workspace you are working in. A personal workspace reads your own rows, a club workspace reads that club's. Nothing crosses.
The club standards check retrieves, the rest do not
Judging a drill against a club methodology means finding the right part of that methodology first. So the standards check runs a retrieval step: your club documents are embedded with text-embedding-3-small, then ranked by a blend of semantic similarity, plain keyword overlap and how authoritative the source is. The top 5 go into the prompt, and findings name the standard they came from. The other seven tools need no retrieval, because the record they read is the one you clicked on.
Club profileAge-band guidanceStandard templatesApproved library items
Where a number matters, the model does not choose it
AVERAGESA progress report's dimension scores are replaced with the averages the server computes from your evaluations before the report is shown
SCORESA standards check's score out of 100 and its verdict are derived from the findings by rule, so the same submission scores the same twice
FIXTURESFixture extraction is instructed never to invent a fixture, fill a gap or guess a kickoff time, and every row is editable before it saves
GAPSWhere the record is too thin to support a claim, the brief is told to say so rather than pad it into a confident paragraph
gpt-5.4 · temperature 0.2 · fixed output schema
04 · The line

It informs. It never authors.

The useful question about AI in a coaching product is not what it can do, it is what it is allowed to finish. Here nothing it writes reaches a player, a parent or the club library without a coach pressing something first, and several things people expect to find here simply do not exist.

What Teloframe does not have
✕There is no chat box and no free-form prompt anywhere in the product
✕No AI picks a team, writes a lineup or decides who plays
✕No AI generates a session plan or recommends a drill
✕The 40 board patterns are hand-authored by people, not generated
✕No video analysis, no clip tagging, no GPS or wearable data
✕Nothing is sent to a parent, a player or the club library automatically
·
The parent summary has two buttons
Generating it and sending it are separate actions. The send only reaches guardians already linked to the player portal, it refuses with a plain message when no family is linked, and it re-checks that what it is sending is the summary the generator produced rather than free text.
·
A photo import asks before it lands
A read it is not confident about will not apply at all: it tells you what it could not make out and stops. A confident read still asks you to confirm, because applying it replaces the players and annotations on the board.
·
Imported fixtures wait in a table
Extraction fills a review grid, not your calendar. Likely duplicates arrive deselected, every field is editable, and nothing is written until you press add. The same table serves CSV and calendar files, which cost no credits on any plan.
·
The standards check is marked advisory
The score renders with an AI estimate label next to it. Findings are a starting list for the human reviewing the submission, who still approves it into the club library or sends it back.
05 · What it costs

Credits, charged at cost.

AI is the one part of the product with a real per-use cost, so it is the one part priced per use. A credit is two cents of model spend. Every plan comes with credits, the balance sits in your account menu, and you can top up without changing plan.

Included with every plan
Free
10 credits
A one-off welcome grant. Enough to try every tool that is not plan-gated.
Pro
100 credits a month
Resets each billing period. Unused credits do not roll over.
Club
500 credits a month
One pool shared across every member of staff in the workspace.
Top-ups · any plan, no upgrade needed
$5.9950 credits$14.99200 credits
Top-up credits sit alongside your included ones and are spent last, so a monthly allowance is never wasted. On Club, both pools are shared by the whole staff.
Held, then charged at cost
A generation reserves the maximum it could cost, then settles at what the model actually used. One credit is two cents of model spend.
A failed run costs nothing
If the model errors or returns a shape that fails validation, the hold is released in full. You are never charged for output you did not get.
A repeat costs nothing
Ask for the same thing on the same unchanged record and the cached result comes back for zero credits. Match debriefs are saved, so reopening one is always free.
One click, one charge
A double click or a second tab is folded into the same generation, so an impatient click cannot bill you twice.
Plan gatesSix of the eight tools work on every plan, including Free: progress report, parent summary, review prep, opponent prep, match debrief and whiteboard photo import. Fixture text extraction needs Pro, because pasting a league email is a convenience and CSV and calendar import stay free. The club standards check needs a Club workspace and library review rights, because there is no club methodology to check against without one.
06 · Privacy

Children's data is in this record. We treat it that way.

These tools read evaluations, notes and match records about named young players. What that means in practice is written down: what leaves the platform, who processes it, how long it is kept, and what is logged every time a tool runs.

Processed, not trained on
Commercial API terms
Prompts and outputs are processed by OpenAI under a commercial API arrangement rather than a consumer product. We do not use your content to train our own models.
Scoped to your workspace
Every loader, every query
Each context loader filters to the workspace you are in and the club it belongs to. Another club's methodology is not reachable from your prompt.
Metered, line by line
Feature · tokens · cost · credits
Every generation records which tool ran, the tokens in and out, the cost, the credits charged, how long it took and whether it was served from cache.
Twelve month retention
Prompts and outputs
AI prompts, outputs and generation records are kept for twelve months. Credit balances and usage counters are kept while the account is open, for billing and quota.
Labelled on screen
Everywhere it appears
Anything a model wrote is marked as generated where it renders, with the time it was generated and the credits it cost. Nothing is passed off as hand-written.
The send is audited
Workspace audit log
Sending a parent summary writes an audit entry naming who sent it, for which player and how many families received it.
Using an AI tool on a player record means that record leaves the platform to be processed. That is your call to make, not ours to make quietly: the full list of who processes what is on the subprocessors page, and section 11 of the privacy policy covers AI features specifically. Every tool is opt-in per click. None of them runs in the background.

AI coaching assistant FAQs

Short answers on which tools exist, what they read, what they cost, and where the coach stays in control.

Is there a chatbot to ask coaching questions?
No. Teloframe has no chat box and no free-form prompt. It has eight buttons, each in the screen that already holds its context: the player file, the team page, the tactics board, the club standards tab. Each one has a fixed job and a fixed output shape, so there is nothing to phrase and nothing to prompt-engineer.
What are the eight AI tools?
Progress report, parent summary and review prep brief in the player file; opponent prep and match debrief on the team page; a standards check in the club workspace; whiteboard photo import on the tactics board; and fixture extraction from pasted text. Six work on every plan including Free, fixture extraction needs Pro, and the standards check needs a Club workspace.
Where does the AI get its information?
From your own record and nothing else. Each tool has a loader that reads named tables in your workspace: evaluations, development goals, match entries and timelines, session themes, and for the standards check your club's written methodology and age-band guidance. There is no web browsing, no scouting feed and no shared corpus across clubs.
Can the AI publish anything on its own?
No. A parent summary is generated and sent as two separate actions, and the send only reaches guardians already linked to the player portal. Imported fixtures land in a review table you edit before saving. A photo import asks before it replaces your board. A standards check is labelled advisory and a human reviewer still decides.
What does an AI generation cost?
Between one and four credits depending on the tool, where a credit is two cents of model spend. Free includes ten welcome credits, Pro includes 100 a month and Club includes 500 a month pooled across staff. Top-ups are 50 credits for $5.99 or 200 for $14.99. A run that fails costs nothing, and asking again for the same unchanged record costs nothing.
Does the AI make selection or tactical decisions?
No. Nothing picks a team, writes a lineup, generates a session plan or recommends a drill. The 40 tactical patterns on the board are hand-authored by people, not generated. Where a number appears in AI output it is computed by the server, not chosen by the model: dimension averages come from your evaluations and a standards score is derived from the findings by rule.
Is our data used to train AI models?
No. Prompts and outputs are processed by OpenAI under a commercial API arrangement rather than a consumer product, and customer content is not used to train our own models. AI prompts, outputs and generation records are kept for twelve months. Every generation is metered and anything a model wrote is labelled as generated where it appears.
The honest version

Eight tools that read your record. Nothing that guesses.

No chat box, no scouting feed, no model writing the coaching judgement. Teloframe's AI turns the evaluations, matches and sessions you already logged into the paragraph you were going to write, then hands it back for you to edit and sign.

Start free →See player development
Teloframe

The Coach OS for football. Plan, run, observe, reflect, develop, one loop, one tool.

Product
Tactics boardDrill editorSession plannerMatch dayPlayer developmentClub workspaceAI assistantExport and sharingAll features
Explore
Tactics librarySession plansDrill libraryFormationsGuidesVideo tutorials
Use cases
Coaches and academiesContent creatorsSession planner softwareTactics board software
Company
PricingFAQContactOpen Teloframe
© 2026 Teloframe. Built for serious staff.PrivacyTermsSafeguardingSecurityDPASubprocessorsCookies All systems normal